U.S. patent application number 13/667456 was filed with the patent office on 2013-03-07 for optimizing the deployment of a workload on a distributed processing system.
This patent application is currently assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION. The applicant listed for this patent is INTERNATIONAL BUSINESS MACHINES CORP. Invention is credited to CHARLES J. ARCHER, MARK G. MEGERIAN, GARY R. RICARD, BRIAN E. SMITH.
Application Number | 20130061238 13/667456 |
Document ID | / |
Family ID | 46491742 |
Filed Date | 2013-03-07 |
United States Patent
Application |
20130061238 |
Kind Code |
A1 |
ARCHER; CHARLES J. ; et
al. |
March 7, 2013 |
OPTIMIZING THE DEPLOYMENT OF A WORKLOAD ON A DISTRIBUTED PROCESSING
SYSTEM
Abstract
Optimizing the deployment of a workload on a distributed
processing system, the distributed processing system having a
plurality of nodes, each node having a plurality of attributes,
including: profiling during operations on the distributed
processing system attributes of the nodes of the distributed
processing system; selecting a workload for deployment on a subset
of the nodes of the distributed processing system; determining
specific resource requirements for the workload to be deployed;
determining a required geometry of the nodes to run the workload;
selecting a set of nodes having attributes that meet the specific
resource requirements and arranged to meet the required geometry;
deploying the workload on the selected nodes.
Inventors: |
ARCHER; CHARLES J.;
(ROCHESTER, MN) ; MEGERIAN; MARK G.; (ROCHESTER,
MN) ; RICARD; GARY R.; (CHATFIELD, MN) ;
SMITH; BRIAN E.; (KNOXVILLE, TN) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
INTERNATIONAL BUSINESS MACHINES CORP; |
Armonk |
NY |
US |
|
|
Assignee: |
INTERNATIONAL BUSINESS MACHINES
CORPORATION
ARMONK
NY
|
Family ID: |
46491742 |
Appl. No.: |
13/667456 |
Filed: |
November 2, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13007905 |
Jan 17, 2011 |
|
|
|
13667456 |
|
|
|
|
Current U.S.
Class: |
718/105 |
Current CPC
Class: |
G06F 9/5066 20130101;
G06F 9/5044 20130101 |
Class at
Publication: |
718/105 |
International
Class: |
G06F 9/46 20060101
G06F009/46 |
Claims
1. A method of optimizing the deployment of a workload on a
distributed processing system, the distributed processing system
having a plurality of nodes, each node having a plurality of
attributes, the method comprising: profiling during operations on
the distributed processing system attributes of the nodes of the
distributed processing system; selecting a workload for deployment
on a subset of the nodes of the distributed processing system;
determining specific resource requirements for the workload to be
deployed; determining a required geometry of the nodes to run the
workload; selecting a set of nodes having attributes that meet the
specific resource requirements and arranged to meet the required
geometry; and deploying the workload on the selected nodes.
2. The method of claim 1 wherein selecting a set of nodes having
attributes that meet the specific resource requirements and
arranged to meet the required geometry further comprises: selecting
a plurality of candidate sets of nodes; assigning to each candidate
set of nodes a score, the score being a representation of the
degree to which the attributes of the nodes of the set meet the
resource requirements of the workload and the geometry requirements
of the workload; and selecting the candidate set of nodes having
the best score.
3. The method of claim 1 wherein profiling during operations on the
distributed processing system attributes of the nodes of the
distributed processing system comprises profiling the attributes of
a set of nodes during a previous run of the workload; and selecting
a set of nodes having attributes that meet the specific resource
requirements and arranged to meet the required geometry further
comprises selecting a set of nodes that are different than those
used in the previous run of the workload; and deploying the
workload on the selected nodes further comprises suggesting the set
of nodes that are different than those used in the previous run of
the workload for the next run of the workload.
4. The method of claim 1 wherein profiling during operations on the
distributed processing system attributes of the nodes of the
distributed processing system further comprises: running a system
exerciser on the distributed processing system, the system
exerciser comprising operations to test the attributes of the
nodes; and recording the resultant performance of the attributes of
the nodes in response to the system exerciser; and selecting a set
of nodes having attributes that meet the specific resource
requirements and arranged to meet the required geometry further
comprises suggesting an initial set of nodes for deploying the
workload.
5. The method of claim 1 wherein determining specific resource
requirements for the workload to be deployed further comprises
receiving specific resource requirements from the user.
6. The method of claim 1 wherein determining specific resource
requirements for the workload to be deployed further comprises
monitoring the consumption of various resources by the workload in
one or more runs of the workload.
7. The method of claim 1 wherein profiling during operations on the
distributed processing system attributes of the nodes of the
distributed processing system further comprises storing in a
database an identification of the nodes and the specific attributes
of the nodes.
8. The method of claim 1 wherein various nodes of the distributed
processing system have different components from one another.
9. The method of claim 1 further comprising identifying in
dependence upon the attributes of the nodes of the distributed
processing system components to be replaced and suggesting the
replacement of the components.
10-25. (canceled)
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application is a continuation application of and claims
priority from U.S. patent application Ser. No. 13/007,905, filed on
Jan. 17, 2011.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The field of the invention is data processing, or, more
specifically, methods, apparatus, and products for optimizing the
deployment of a workload on a distributed processing system.
[0004] 2. Description of Related Art
[0005] The development of the EDVAC computer system of 1948 is
often cited as the beginning of the computer era. Since that time,
computer systems have evolved into extremely complicated devices.
Today's computers are much more sophisticated than early systems
such as the EDVAC. Computer systems typically include a combination
of hardware and software components, application programs,
operating systems, processors, buses, memory, input/output devices,
and so on. As advances in semiconductor processing and computer
architecture push the performance of the computer higher and
higher, more sophisticated computer software has evolved to take
advantage of the higher performance of the hardware, resulting in
computer systems today that are much more powerful than just a few
years ago.
[0006] Next generation supercomputers and distributed compute
platforms contain many execution nodes, many of which have
different components and different processing capabilities. As
such, one execution node may be able to execute particular
workloads in a more efficient manner than another execution node
because of the differences between the two execution nodes.
Computing workloads may therefore be more efficiently executed by
scheduling and assigning the workloads in a way to better utilize
the resources in a particular system.
SUMMARY OF THE INVENTION
[0007] Methods, apparatus, and products for optimizing the
deployment of a workload on a distributed processing system, the
distributed processing system having a plurality of nodes, each
node having a plurality of attributes, including: profiling during
operations on the distributed processing system attributes of the
nodes of the distributed processing system; selecting a workload
for deployment on a subset of the nodes of the distributed
processing system; determining specific resource requirements for
the workload to be deployed; determining a required geometry of the
nodes to run the workload; selecting a set of nodes having
attributes that meet the specific resource requirements and
arranged to meet the required geometry; and deploying the workload
on the selected nodes.
[0008] The foregoing and other objects, features and advantages of
the invention will be apparent from the following more particular
descriptions of exemplary embodiments of the invention as
illustrated in the accompanying drawings wherein like reference
numbers generally represent like parts of exemplary embodiments of
the invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 sets forth example apparatus for optimizing the
deployment of a workload on a distributed processing system
according to embodiments of the present invention.
[0010] FIG. 2 sets forth a block diagram of an example compute node
useful in a parallel computer capable of optimizing the deployment
of a workload on a distributed processing system according to
embodiments of the present invention.
[0011] FIG. 3A sets forth a block diagram of an example
Point-To-Point Adapter useful in optimizing the deployment of a
workload on a distributed processing system according to
embodiments of the present invention.
[0012] FIG. 3B sets forth a block diagram of an example Global
Combining Network Adapter useful in optimizing the deployment of a
workload on a distributed processing system according to
embodiments of the present invention.
[0013] FIG. 4 sets forth a line drawing illustrating an example
data communications network optimized for optimizing the deployment
of a workload on a distributed processing system according to
embodiments of the present invention.
[0014] FIG. 5 sets forth a line drawing illustrating an example
global combining network useful in systems capable of optimizing
the deployment of a workload on a distributed processing system
according to embodiments of the present invention.
[0015] FIG. 6 sets forth a flow chart illustrating an exemplary
method for optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present
invention.
[0016] FIG. 7 sets forth a flow chart illustrating an exemplary
method for optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present
invention.
[0017] FIG. 8 sets forth a flow chart illustrating an exemplary
method for optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present
invention.
[0018] FIG. 9 sets forth a flow chart illustrating an exemplary
method for optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present
invention.
DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS
[0019] Exemplary methods, apparatus, and products for optimizing
the deployment of a workload on a distributed processing system in
accordance with the present invention are described with reference
to the accompanying drawings, beginning with FIG. 1. FIG. 1 sets
forth example apparatus for optimizing the deployment of a workload
on a distributed processing system according to embodiments of the
present invention. The apparatus of FIG. 1 includes a parallel
computer (100), non-volatile memory for the computer in the form of
a data storage device (118), an output device for the computer in
the form of a printer (120), and an input/output device for the
computer in the form of a computer terminal (122). The parallel
computer (100) in the example of FIG. 1 includes a plurality of
compute nodes (102). The compute nodes (102) are coupled for data
communications by several independent data communications networks
including a high speed Ethernet network (174), a Joint Test Action
Group (`JTAG`) network (104), a global combining network (106)
which is optimized for collective operations using a binary tree
network topology, and a point-to-point network (108), which is
optimized for point-to-point operations using a torus network
topology. The global combining network (106) is a data
communications network that includes data communications links
connected to the compute nodes (102) so as to organize the compute
nodes (102) as a binary tree. Each data communications network is
implemented with data communications links among the compute nodes
(102). The data communications links provide data communications
for parallel operations among the compute nodes (102) of the
parallel computer (100).
[0020] The compute nodes (102) of the parallel computer (100) are
organized into at least one operational group (132) of compute
nodes for collective parallel operations on the parallel computer
(100). Each operational group (132) of compute nodes is the set of
compute nodes upon which a collective parallel operation executes.
Each compute node in the operational group (132) is assigned a
unique rank that identifies the particular compute node in the
operational group (132). Collective operations are implemented with
data communications among the compute nodes of a operational group.
Collective operations are those functions that involve all the
compute nodes of an operational group (132). A collective operation
is an operation, a message-passing computer program instruction
that is executed simultaneously, that is, at approximately the same
time, by all the compute nodes in an operational group (132) of
compute nodes. Such an operational group (132) may include all the
compute nodes (102) in a parallel computer (100) or a subset all
the compute nodes (102). Collective operations are often built
around point-to-point operations. A collective operation requires
that all processes on all compute nodes within an operational group
(132) call the same collective operation with matching arguments. A
`broadcast` is an example of a collective operation for moving data
among compute nodes of a operational group. A `reduce` operation is
an example of a collective operation that executes arithmetic or
logical functions on data distributed among the compute nodes of a
operational group (132). An operational group (132) may be
implemented as, for example, an MPI `communicator.`
[0021] `MPI` refers to `Message Passing Interface,` a prior art
parallel communications library, a module of computer program
instructions for data communications on parallel computers.
Examples of prior-art parallel communications libraries that may be
improved for performing an allreduce operation using shared memory
according to embodiments of the present invention include MPI and
the `Parallel Virtual Machine` (`PVM`) library. PVM was developed
by the University of Tennessee, The Oak Ridge National Laboratory
and Emory University. MPI is promulgated by the MPI Forum, an open
group with representatives from many organizations that define and
maintain the MPI standard. MPI at the time of this writing is a de
facto standard for communication among compute nodes running a
parallel program on a distributed memory parallel computer. This
specification sometimes uses MPI terminology for ease of
explanation, although the use of MPI as such is not a requirement
or limitation of the present invention.
[0022] Some collective operations have a single originating or
receiving process running on a particular compute node in an
operational group (132). For example, in a `broadcast` collective
operation, the process on the compute node that distributes the
data to all the other compute nodes is an originating process. In a
`gather` operation, for example, the process on the compute node
that received all the data from the other compute nodes is a
receiving process. The compute node on which such an originating or
receiving process runs is referred to as a logical root.
[0023] Most collective operations are variations or combinations of
four basic operations: broadcast, gather, scatter, and reduce. The
interfaces for these collective operations are defined in the MPI
standards promulgated by the MPI Forum. Algorithms for executing
collective operations, however, are not defined in the MPI
standards. In a broadcast operation, all processes specify the same
root process, whose buffer contents will be sent. Processes other
than the root specify receive buffers. After the operation, all
buffers contain the message from the root process.
[0024] A scatter operation, like the broadcast operation, is also a
one-to-many collective operation. In a scatter operation, the
logical root divides data on the root into segments and distributes
a different segment to each compute node in the operational group
(132). In scatter operation, all processes typically specify the
same receive count. The send arguments are only significant to the
root process, whose buffer actually contains sendcount * N elements
of a given datatype, where N is the number of processes in the
given group of compute nodes. The send buffer is divided and
dispersed to all processes (including the process on the logical
root). Each compute node is assigned a sequential identifier termed
a `rank.` After the operation, the root has sent sendcount data
elements to each process in increasing rank order. Rank 0 receives
the first sendcount data elements from the send buffer. Rank 1
receives the second sendcount data elements from the send buffer,
and so on.
[0025] A gather operation is a many-to-one collective operation
that is a complete reverse of the description of the scatter
operation. That is, a gather is a many-to-one collective operation
in which elements of a datatype are gathered from the ranked
compute nodes into a receive buffer in a root node.
[0026] A reduction operation is also a many-to-one collective
operation that includes an arithmetic or logical function performed
on two data elements. All processes specify the same `count` and
the same arithmetic or logical function. After the reduction, all
processes have sent count data elements from computer node send
buffers to the root process. In a reduction operation, data
elements from corresponding send buffer locations are combined
pair-wise by arithmetic or logical operations to yield a single
corresponding element in the root process' receive buffer.
Application specific reduction operations can be defined at
runtime. Parallel communications libraries may support predefined
operations. MPI, for example, provides the following pre-defined
reduction operations:
TABLE-US-00001 MPI_MAX maximum MPI_MIN minimum MPI_SUM sum MPI_PROD
product MPI_LAND logical and MPI_BAND bitwise and MPI_LOR logical
or MPI_BOR bitwise or MPI_LXOR logical exclusive or MPI_BXOR
bitwise exclusive or
[0027] In addition to compute nodes, the parallel computer (100)
includes input/output (`I/O`) nodes (110, 114) coupled to compute
nodes (102) through the global combining network (106). The compute
nodes (102) in the parallel computer (100) may be partitioned into
processing sets such that each compute node in a processing set is
connected for data communications to the same I/O node. Each
processing set, therefore, is composed of one I/O node and a subset
of compute nodes (102). The ratio between the number of compute
nodes to the number of I/O nodes in the entire system typically
depends on the hardware configuration for the parallel computer
(102). For example, in some configurations, each processing set may
be composed of eight compute nodes and one I/O node. In some other
configurations, each processing set may be composed of sixty-four
compute nodes and one I/O node. Such example are for explanation
only, however, and not for limitation. Each I/O node provides I/O
services between compute nodes (102) of its processing set and a
set of I/O devices. In the example of FIG. 1, the I/O nodes (110,
114) are connected for data communications I/O devices (118, 120,
122) through local area network (LAN') (130) implemented using
high-speed Ethernet.
[0028] The parallel computer (100) of FIG. 1 also includes a
service node (116) coupled to the compute nodes through one of the
networks (104). Service node (116) provides services common to
pluralities of compute nodes, administering the configuration of
compute nodes, loading programs into the compute nodes, starting
program execution on the compute nodes, retrieving results of
program operations on the computer nodes, and so on. Service node
(116) runs a service application (124) and communicates with users
(128) through a service application interface (126) that runs on
computer terminal (122).
[0029] In the example of FIG. 1, the service node (116) includes a
deployment module (103) configured to optimize the deployment of a
workload on a distributed processing system according to
embodiments of the present invention. In the example of FIG. 1,
each of the compute nodes (102) in the parallel computer (100) is
characterized by a plurality of attributes that describe
characteristics of each compute node (102). For example, attributes
that describe characteristics of the compute nodes (102) may
include, for example, the speed of a CPU on the a particular
compute node, the amount of memory on a particular compute node,
the location of particular compute node in the parallel computer
(100), and other attributes describing the characteristics of
components within particular compute node. In addition, a
particular compute node may also characterized by a plurality of
attributes that describe characteristics of a particular compute
node in the sense that the attributes describe performance-related
characteristics of a particular compute node such as, for example,
the speed at which a particular compute node can execute floating
point operations, the speed at which a particular compute node can
transmit a message to another compute node, the speed at which a
particular compute node can execute various I/O operations, and so
on.
[0030] In the example of FIG. 1, the deployment module (103)
optimizes the deployment of a workload by profiling during
operations on the distributed processing system attributes of the
nodes of the parallel computer (100). In the example of FIG. 1,
profiling attributes of the compute nodes (102) may be carried out,
for example, by recording performance metrics of the compute nodes
(102). Such performance metrics may be recorded and compiled such
that a profile regarding the performance attributes of the compute
nodes (102) can be created. For example, performance metrics may be
recorded and compiled such that a profile regarding the performance
attributes of the compute nodes (102) can be created that details
the amount of time a particular compute node took to carry out a
floating point operation, the amount of time a particular compute
node took to carry out an I/O operation, the amount of time a
particular compute node took to carry out a data communications
operation, and other performance metrics can be recorded for the
purposes of profiling attributes of the compute nodes (102) of the
parallel computer (100).
[0031] In the example of FIG. 1, the deployment module (103) also
optimizes the deployment of a workload by selecting a workload for
deployment on a subset of the compute nodes (102) of the parallel
computer (100). In the example of FIG. 1, the workload represents a
series of computer program instructions that are to be executed.
Workloads may be characterized, for example, by the amount of
processor cycles that are required to execute the workload, the
amount of memory that is required to execute the workload, the
amount of a particular type of operations are required to execute
the workload, and so on. In the example of FIG. 1, a workload may
be selected for deployment on a subset of the compute nodes (102)
of the parallel computer (100) based on, for example, a
shortest-time-to-completion scheduling algorithm, a
first-in-first-out scheduling algorithm, the availability of a
particular type of system resource, a priority associated with the
workload, and so on.
[0032] In the example of FIG. 1, the deployment module (103) also
optimizes the deployment of a workload by determining specific
resource requirements for the workload to be deployed. In the
example of FIG. 1, determining specific resource requirements for
the workload to be deployed may be carried out, for example, by
examining the computer program instructions that make up the
workload. The computer program instructions that make up the
workload may be examined to determine, for example, the number of
floating point instructions in the workload, the amount of data
communications operations in the workload, the amount of I/O
operations in the workload, the number of message passing
operations in the workload, and so on. Determining specific
resource requirements for the workload to be deployed may therefore
be carried out by determining the nature and amount of system
resources that are needed to execute each of the component parts of
the workload.
[0033] In the example of FIG. 1, the deployment module (103) also
optimizes the deployment of a workload by determining a required
geometry of the compute nodes (102) to run the workload. In the
example of FIG. 1, a required geometry of the compute nodes (102)
to run the workload may be determined, for example, based on the
computer program instructions that make up the workload. For
example, a particular workload may include a collective operation.
In such an example, the collective operation requires that the
compute nodes (102) are organized as a tree. The required geometry
of the compute nodes (102) to run the workload in such an example
is therefore a tree geometry. Alternatively, a particular workload
may require a high number of data communications operations such
that a geometry of compute nodes (102) in which the compute nodes
(102) are within close physical proximity of each other may be
preferred so as to avoid data communications between compute nodes
(102) over long physical distances.
[0034] In the example of FIG. 1, the deployment module (103) also
optimizes the deployment of a workload by selecting a set of
compute nodes (102) having attributes that meet the specific
resource requirements and arranged to meet the required geometry
required by a workload. In the example of FIG. 1, selecting a set
of compute nodes (102) having attributes that meet the specific
resource requirements and arranged to meet the required geometry
may be carried out, for example, by comparing the attributes of a
particular set of compute nodes (102) to the resource requirements
and required geometry for a workload to determine a best match.
Determining a best match may include prioritizing specific resource
requirements of the workload, determining a score for a candidate
set of compute nodes (102), and so on.
[0035] In the example of FIG. 1, the deployment module (103) also
optimizes the deployment of a workload by deploying the workload on
the selected compute nodes (102). In the example of FIG. 1,
deploying the workload on the selected compute nodes (102) may be
carried out, for example, by sending the workload, or a portion
thereof, to an execution queue on the selected compute nodes (102),
assigning the workload for execution on processors of the selected
compute nodes (102), and so on.
[0036] The arrangement of nodes, networks, and I/O devices making
up the example apparatus illustrated in FIG. 1 are for explanation
only, not for limitation of the present invention. Apparatus
capable of optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present invention
may include additional nodes, networks, devices, and architectures,
not shown in FIG. 1, as will occur to those of skill in the art.
The parallel computer (100) in the example of FIG. 1 includes
sixteen compute nodes (102); parallel computers capable of
optimizing the deployment of a workload on a distributed processing
system according to embodiments of the present invention sometimes
include thousands of compute nodes. In addition to Ethernet (174)
and JTAG (104), networks in such data processing systems may
support many data communications protocols including for example
TCP (Transmission Control Protocol), IP (Internet Protocol), and
others as will occur to those of skill in the art. Various
embodiments of the present invention may be implemented on a
variety of hardware platforms in addition to those illustrated in
FIG. 1.
[0037] Optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present invention
is generally implemented on a parallel computer that includes a
plurality of compute nodes organized for collective operations
through at least one data communications network. In fact, such
parallel computers may include thousands of such compute nodes.
Each compute node is in turn itself a kind of computer composed of
one or more computer processing cores, its own computer memory, and
its own input/output adapters. For further explanation, therefore,
FIG. 2 sets forth a block diagram of an example compute node (102)
useful in a parallel computer capable of optimizing the deployment
of a workload on a distributed processing system according to
embodiments of the present invention. The compute node (102) of
FIG. 2 includes a plurality of processing cores (165) as well as
RAM (156). The processing cores (165) of FIG. 2 may be configured
on one or more integrated circuit dies. Processing cores (165) are
connected to RAM (156) through a high-speed memory bus (155) and
through a bus adapter (194) and an extension bus (168) to other
components of the compute node. Stored in RAM (156) is an
application program (159), a module of computer program
instructions that carries out parallel, user-level data processing
using parallel algorithms.
[0038] Also stored RAM (156) is a parallel communications library
(161), a library of computer program instructions that carry out
parallel communications among compute nodes, including
point-to-point operations as well as collective operations.
Application program (159) executes collective operations by calling
software routines in parallel communications library (161). A
library of parallel communications routines may be developed from
scratch for use in systems according to embodiments of the present
invention, using a traditional programming language such as the C
programming language, and using traditional programming methods to
write parallel communications routines that send and receive data
among nodes on two independent data communications networks.
Alternatively, existing prior art libraries may be improved to
operate according to embodiments of the present invention. Examples
of prior-art parallel communications libraries include the `Message
Passing Interface` (`MPI`) library and the `Parallel Virtual
Machine` (`PVM`) library.
[0039] Also stored in RAM (156) is a compute node operating system
(162), a module of computer program instructions and routines for
an application program's access to other resources of the compute
node. It is typical for an application program (159) and parallel
communications library (161) in a compute node (102) of a parallel
computer to run a single thread of execution with no user login and
no security issues because the thread is entitled to complete
access to all resources of the compute node (102). Operating
systems that may usefully be improved, simplified, for use in a
compute node include UNIX.TM., Linux.TM., Microsoft XP.TM.,
AIX.TM., IBM's i5/OS.TM., and others as will occur to those of
skill in the art. In the example of FIG. 2, the compute node
operating system (162) of FIG. 2 may be capable of supporting one
or more virtual machines. The compute node operating system (162)
of FIG. 2 may therefore include virtual machine management
components such as, for example, a hypervisor or other module of
automated computing machinery capable of supporting one or more
virtual machines.
[0040] The example compute node (102) of FIG. 2 includes several
communications adapters (172, 176, 180, 188) for implementing data
communications with other nodes of a parallel computer. Such data
communications may be carried out serially through RS-232
connections, through external buses such as USB, through data
communications networks such as IP networks, and in other ways as
will occur to those of skill in the art. Communications adapters
implement the hardware level of data communications through which
one computer sends data communications to another computer,
directly or through a network. Examples of communications adapters
useful in apparatus that page memory from RAM to backing storage in
a parallel computer include modems for wired communications,
Ethernet (IEEE 802.3) adapters for wired network communications,
and 802.11b adapters for wireless network communications.
[0041] The data communications adapters in the example of FIG. 2
include a Gigabit Ethernet adapter (172) that couples example
compute node (102) for data communications to a Gigabit Ethernet
(174). Gigabit Ethernet is a network transmission standard, defined
in the IEEE 802.3 standard, that provides a data rate of 1 billion
bits per second (one gigabit). Gigabit Ethernet is a variant of
Ethernet that operates over multimode fiber optic cable, single
mode fiber optic cable, or unshielded twisted pair.
[0042] The data communications adapters in the example of FIG. 2
include a JTAG Slave circuit (176) that couples example compute
node (102) for data communications to a JTAG Master circuit (178).
JTAG is the usual name used for the IEEE 1149.1 standard entitled
Standard Test Access Port and Boundary-Scan Architecture for test
access ports used for testing printed circuit boards using boundary
scan. JTAG is so widely adapted that, at this time, boundary scan
is more or less synonymous with JTAG. JTAG is used not only for
printed circuit boards, but also for conducting boundary scans of
integrated circuits, and is also useful as a mechanism for
debugging embedded systems, providing a convenient "back door" into
the system. The example compute node of FIG. 2 may be all three of
these: It typically includes one or more integrated circuits
installed on a printed circuit board and may be implemented as an
embedded system having its own processing core, its own memory, and
its own I/O capability. JTAG boundary scans through JTAG Slave
(176) may efficiently configure processing core registers and
memory in compute node (102) for use in dynamically reassigning a
connected node to a block of compute nodes for paging memory from
RAM to backing storage in a parallel computer according to
embodiments of the present invention.
[0043] The data communications adapters in the example of FIG. 2
include a Point-To-Point Network Adapter (180) that couples example
compute node (102) for data communications to a network (108) that
is optimal for point-to-point message passing operations such as,
for example, a network configured as a three-dimensional torus or
mesh. The Point-To-Point Adapter (180) provides data communications
in six directions on three communications axes, x, y, and z,
through six bidirectional links: +x (181), -x (182), +y (183), -y
(184), +z (185), and -z (186).
[0044] The data communications adapters in the example of FIG. 2
include a Global Combining Network Adapter (188) that couples
example compute node (102) for data communications to a global
combining network (106) that is optimal for collective message
passing operations such as, for example, a network configured as a
binary tree. The Global Combining Network Adapter (188) provides
data communications through three bidirectional links for each
global combining network (106) that the Global Combining Network
Adapter (188) supports. In the example of FIG. 2, the Global
Combining Network Adapter (188) provides data communications
through three bidirectional links for global combining network
(106): two to children nodes (190) and one to a parent node
(192).
[0045] The example compute node (102) includes multiple arithmetic
logic units (`ALUs`). Each processing core (165) includes an ALU
(166), and a separate ALU (170) is dedicated to the exclusive use
of the Global Combining Network Adapter (188) for use in performing
the arithmetic and logical functions of reduction operations,
including an allreduce operation. Computer program instructions of
a reduction routine in a parallel communications library (161) may
latch an instruction for an arithmetic or logical function into an
instruction register (169). When the arithmetic or logical function
of a reduction operation is a `sum` or a `logical OR,` for example,
the collective operations adapter (188) may execute the arithmetic
or logical operation by use of the ALU (166) in the processing core
(165) or, typically much faster, by use of the dedicated ALU (170)
using data provided by the nodes (190, 192) on the global combining
network (106) and data provided by processing cores (165) on the
compute node (102).
[0046] Often when performing arithmetic operations in the global
combining network adapter (188), however, the global combining
network adapter (188) only serves to combine data received from the
children nodes (190) and pass the result up the network (106) to
the parent node (192). Similarly, the global combining network
adapter (188) may only serve to transmit data received from the
parent node (192) and pass the data down the network (106) to the
children nodes (190). That is, none of the processing cores (165)
on the compute node (102) contribute data that alters the output of
ALU (170), which is then passed up or down the global combining
network (106). Because the ALU (170) typically does not output any
data onto the network (106) until the ALU (170) receives input from
one of the processing cores (165), a processing core (165) may
inject the identity element into the dedicated ALU (170) for the
particular arithmetic operation being perform in the ALU (170) in
order to prevent alteration of the output of the ALU (170).
Injecting the identity element into the ALU, however, often
consumes numerous processing cycles. To further enhance performance
in such cases, the example compute node (102) includes dedicated
hardware (171) for injecting identity elements into the ALU (170)
to reduce the amount of processing core resources required to
prevent alteration of the ALU output. The dedicated hardware (171)
injects an identity element that corresponds to the particular
arithmetic operation performed by the ALU. For example, when the
global combining network adapter (188) performs a bitwise OR on the
data received from the children nodes (190), dedicated hardware
(171) may inject zeros into the ALU (170) to improve performance
throughout the global combining network (106).
[0047] For further explanation, FIG. 3A sets forth a block diagram
of an example Point-To-Point Adapter (180) useful in optimizing the
deployment of a workload on a distributed processing system
according to embodiments of the present invention. The
Point-To-Point Adapter (180) is designed for use in a data
communications network optimized for point-to-point operations, a
network that organizes compute nodes in a three-dimensional torus
or mesh. The Point-To-Point Adapter (180) in the example of FIG. 3A
provides data communication along an x-axis through four
unidirectional data communications links, to and from the next node
in the -x direction (182) and to and from the next node in the +x
direction (181). The Point-To-Point Adapter (180) of FIG. 3A also
provides data communication along a y-axis through four
unidirectional data communications links, to and from the next node
in the -y direction (184) and to and from the next node in the +y
direction (183). The Point-To-Point Adapter (180) of FIG. 3A also
provides data communication along a z-axis through four
unidirectional data communications links, to and from the next node
in the -z direction (186) and to and from the next node in the +z
direction (185).
[0048] For further explanation, FIG. 3B sets forth a block diagram
of an example Global Combining Network Adapter (188) useful in
optimizing the deployment of a workload on a distributed processing
system according to embodiments of the present invention. The
Global Combining Network Adapter (188) is designed for use in a
network optimized for collective operations, a network that
organizes compute nodes of a parallel computer in a binary tree.
The Global Combining Network Adapter (188) in the example of FIG.
3B provides data communication to and from children nodes of a
global combining network through four unidirectional data
communications links (190), and also provides data communication to
and from a parent node of the global combining network through two
unidirectional data communications links (192).
[0049] For further explanation, FIG. 4 sets forth a line drawing
illustrating an example data communications network (108) optimized
for optimizing the deployment of a workload on a distributed
processing system according to embodiments of the present
invention. In the example of FIG. 4, dots represent compute nodes
(102) of a parallel computer, and the dotted lines between the dots
represent data communications links (103) between compute nodes.
The data communications links are implemented with point-to-point
data communications adapters similar to the one illustrated for
example in FIG. 3A, with data communications links on three axis,
x, y, and z, and to and fro in six directions +x (181), -x (182),
+y (183), -y (184), +z (185), and -z (186). The links and compute
nodes are organized by this data communications network optimized
for point-to-point operations into a three dimensional mesh (105).
The mesh (105) has wrap-around links on each axis that connect the
outermost compute nodes in the mesh (105) on opposite sides of the
mesh (105). These wrap-around links form a torus (107). Each
compute node in the torus has a location in the torus that is
uniquely specified by a set of x, y, z coordinates. Readers will
note that the wrap-around links in the y and z directions have been
omitted for clarity, but are configured in a similar manner to the
wrap-around link illustrated in the x direction. For clarity of
explanation, the data communications network of FIG. 4 is
illustrated with only 27 compute nodes, but readers will recognize
that a data communications network optimized for point-to-point
operations for use in optimizing the deployment of a workload on a
distributed processing system in accordance with embodiments of the
present invention may contain only a few compute nodes or may
contain thousands of compute nodes. For ease of explanation, the
data communications network of FIG. 4 is illustrated with only
three dimensions, but readers will recognize that a data
communications network optimized for point-to-point operations for
use in optimizing the deployment of a workload on a distributed
processing system in accordance with embodiments of the present
invention may in facet be implemented in two dimensions, four
dimensions, five dimensions, and so on. Several supercomputers now
use five dimensional mesh or torus networks, including, for
example, IBM's Blue Gene Q.TM..
[0050] For further explanation, FIG. 5 sets forth a line drawing
illustrating an example global combining network (106) useful in
systems capable of optimizing the deployment of a workload on a
distributed processing system according to embodiments of the
present invention. The example data communications network of FIG.
5 includes data communications links (103) connected to the compute
nodes so as to organize the compute nodes as a tree. In the example
of FIG. 5, dots represent compute nodes (102) of a parallel
computer, and the dotted lines (103) between the dots represent
data communications links between compute nodes. The data
communications links are implemented with global combining network
adapters similar to the one illustrated for example in FIG. 3B,
with each node typically providing data communications to and from
two children nodes and data communications to and from a parent
node, with some exceptions. Nodes in the global combining network
(106) may be characterized as a physical root node (202), branch
nodes (204), and leaf nodes (206). The physical root (202) has two
children but no parent and is so called because the physical root
node (202) is the node physically configured at the top of the
binary tree. The leaf nodes (206) each has a parent, but leaf nodes
have no children. The branch nodes (204) each has both a parent and
two children. The links and compute nodes are thereby organized by
this data communications network optimized for collective
operations into a binary tree (106). For clarity of explanation,
the data communications network of FIG. 5 is illustrated with only
31 compute nodes, but readers will recognize that a global
combining network (106) optimized for collective operations for use
in optimizing the deployment of a workload on a distributed
processing system in accordance with embodiments of the present
invention may contain only a few compute nodes or may contain
thousands of compute nodes.
[0051] In the example of FIG. 5, each node in the tree is assigned
a unit identifier referred to as a `rank` (250). The rank actually
identifies a task or process that is executing a parallel operation
according to embodiments of the present invention. Using the rank
to identify a node assumes that only one such task is executing on
each node. To the extent that more than one participating task
executes on a single node, the rank identifies the task as such
rather than the node. A rank uniquely identifies a task's location
in the tree network for use in both point-to-point and collective
operations in the tree network. The ranks in this example are
assigned as integers beginning with 0 assigned to the root tasks or
root node (202), 1 assigned to the first node in the second layer
of the tree, 2 assigned to the second node in the second layer of
the tree, 3 assigned to the first node in the third layer of the
tree, 4 assigned to the second node in the third layer of the tree,
and so on. For ease of illustration, only the ranks of the first
three layers of the tree are shown here, but all compute nodes in
the tree network are assigned a unique rank.
[0052] For further explanation, FIG. 6 sets forth a flow chart
illustrating an exemplary method for optimizing the deployment of a
workload on a distributed processing system (601) according to
embodiments of the present invention. In the example of FIG. 6, the
distributed processing system (601) includes a plurality of nodes
(602a-602e). In the example of FIG. 6, each node (602a-602e) is
characterized by a plurality of attributes (606) that describe
characteristics of each node (602a-602e). For example, attributes
(606) that describe characteristics of a node (602a-602e) may
include, for example, the speed of a CPU on the node (602a-602e),
the amount of memory on the node (602a-602e), the location of the
node (602a-602e) in the distributed processing system (601), and
other attributes describing the characteristics of the node
(602a-602e). In addition, each node (602a-602e) may also
characterized by a plurality of attributes (606) that describe
characteristics of a node (602a-602e) in the sense that the
attributes (606) describe performance-related characteristics of
the node (602a-602e) such as, for example, the speed at which a
node (602a-602e) can execute floating point operations, the speed
at which a node (602a-602e) can transmit a message to another node
(602a-602e), the speed at which a node can execute various I/O
operations, and so on. In the example of FIG. 6, various nodes
(602a-602e) of the distributed processing system (601) may have
different components from one another. For example, some nodes may
have local memory while other nodes do not have local memory, some
nodes may have more or different processors that other nodes, some
processor may have different data communications adapters than
other nodes, and so on.
[0053] The example of FIG. 6 includes profiling (604) during
operations on a distributed processing system (602a) attributes
(606) of the nodes (602a-602e) of the distributed processing system
(601). In the example of FIG. 6, profiling (604) attributes (606)
of the nodes (602a-602e) may be carried out, for example, by
recording performance metrics of the nodes (602a-602e). Such
performance metrics may be recorded and compiled such that a
profile regarding the performance attributes of the nodes
(602a-602e) can be created. For example, performance metrics may be
recorded and compiled such that a profile regarding the performance
attributes of the nodes (602a-602e) can be created that details the
amount of time a node (602a-602e) took to carry out a floating
point operation, the amount of time a node (602a-602e) took to
carry out an I/O operation, the amount of time a node (602a-602e)
took to carry out a data communications operation, and other
performance metrics can be recorded for the purposes of profiling
(604) attributes (606) of the nodes (602a-602e) of the distributed
processing system (601).
[0054] The example of FIG. 6 also includes selecting (608) a
workload (612a) for deployment on a subset (622) of the nodes
(602a-602e) of the distributed processing system (601). In the
example of FIG. 6, the workload (612a) represents a series of
computer program instructions that are to be executed. Workloads
(612a-612c) may be characterized, for example, by the amount of
processor cycles that are required to execute the workload
(612a-612c), the amount of memory that is required to execute the
workload (612a-612c), the amount of a particular type of operations
are required to execute the workload (612a-612c), and so on. In the
example of FIG. 6, a workload (612a) may be selected (608) for
deployment on a subset (622) of the nodes (602a-602e) of the
distributed processing system (601) based on, for example, a
shortest-time-to-completion scheduling algorithm, a
first-in-first-out scheduling algorithm, the availability of a
particular type of system resource, a priority associated with the
workload (612a), and so on. In the example of FIG. 6, workloads
(612a-612c) may be stored in a workload queue (610) and removed
from the workload queue (610) as each workload (612a-612c) is
selected for deployment.
[0055] The example of FIG. 6 also includes determining (614)
specific resource requirements for the workload (612a) to be
deployed. In the example of FIG. 6, determining (614) specific
resource requirements for the workload (612a) to be deployed may be
carried out, for example, by examining the computer program
instructions that make up the workload (612a). The computer program
instructions that make up the workload (612a) may be examined to
determine, for example, the number of floating point instructions
in the workload (612a), the amount of data communications
operations in the workload (612a), the amount of I/O operations in
the workload (612a), the number of message passing operations in
the workload (612a), and so on. Determining (614) specific resource
requirements for the workload (612a) to be deployed may therefore
be carried out by determining the nature and amount of system
resources that are needed to execute each of the component parts of
the workload (612a).
[0056] The example of FIG. 6 also includes determining (616) a
required geometry of the nodes (602a-602e) to run the workload
(612a). In the example of FIG. 6, a required geometry of the nodes
(602a-602e) to run the workload (612a) may be determined (616), for
example, based on the computer program instructions that make up
the workload (612a). For example, a particular workload (612a) may
include a collective operation as described above. In such an
example, the collective operation requires that the nodes
(602a-602e) are organized as a tree. The required geometry of the
nodes (602a-602e) to run the workload (612a) in such an example is
therefore a tree geometry. Alternatively, a particular workload
(612a) may require a high number of data communications operations
such that a geometry of nodes (602a-602e) in which the nodes
(602a-602e) are within close physical proximity of each other may
be preferred so as to avoid data communications between nodes
(602a-602e) over long physical distances.
[0057] The example of FIG. 6 also includes selecting (618) a set of
nodes having attributes that meet the specific resource
requirements and arranged to meet the required geometry. In the
example of FIG. 6, selecting (618) a set of nodes having attributes
that meet the specific resource requirements and arranged to meet
the required geometry may be carried out, for example, by comparing
the attributes (606) of a particular set of nodes (602a-602e) to
the resource requirements and required geometry for a workload
(612a) to determine a best match. Determining a best match may
include prioritizing specific resource requirements of the workload
(612a), determining a score for a candidate set of nodes
(602a-602e), and so on.
[0058] The example of FIG. 6 also includes deploying (620) the
workload (612a) on the selected nodes. In the example of FIG. 6,
deploying (620) the workload (612a) on the selected nodes may be
carried out, for example, by sending the workload (612a), or a
portion thereof, to an execution queue on the selected nodes,
assigning the workload (612a) for execution on processors of the
selected nodes, and so on.
[0059] For further explanation, FIG. 7 sets forth a flow chart
illustrating a further exemplary method for optimizing the
deployment of a workload on a distributed processing system (601)
according to embodiments of the present invention. The example of
FIG. 7 is similar to the example of FIG. 6 as it also includes:
[0060] profiling (604) during operations on the distributed
processing system (601) attributes (606) of the nodes (602a-602e)
of the distributed processing system (601), [0061] selecting (608)
a workload (612a) for deployment on a subset (622) of the nodes
(602a-602e) of the distributed processing system (601), [0062]
determining (614) specific resource requirements for the workload
(612a) to be deployed, [0063] determining (616) a required geometry
of the nodes (602a-602e) to run the workload (612a), [0064]
selecting (618) a set of nodes (602a-602e) having attributes that
meet the specific resource requirements and arranged to meet the
required geometry, and [0065] deploying (620) the workload (612a)
on the selected nodes (602a-602e)
[0066] In the example of FIG. 7, determining (614) specific
resource requirements for the workload (612a) to be deployed can
include receiving (702) specific resource requirements from a user.
In the example of FIG. 7, receiving (702) specific resource
requirements from the user can include, for example, receiving
information from a user indicating the number of nodes (602a-602e)
upon which a workload (612a) should run, the amount of memory
needed for executing the workload (612a), and so on. In the example
of FIG. 7, receiving (702) specific resource requirements from the
user can may also include receiving information from a user
indicating a prioritization of system processing capabilities. For
example, a user may indicate that processing I/O operations should
be prioritized over data communications operations, such that nodes
(602a-602e) with high I/O performance will be favored for selection
over nodes that perform data communications operations
efficiently.
[0067] In the example of FIG. 7, determining (614) specific
resource requirements for the workload (612a) to be deployed can
alternatively include monitoring (704) the consumption of various
resources by the workload (612a) in one or more runs of the
workload (612a). In the example of FIG. 7, monitoring (704) the
consumption of various resources by the workload (612a) may
include, for example, monitoring the amount of memory utilized
during the execution of the workload (612a), monitoring processor
usage during the execution of the workload (612a), monitoring the
number of times a communications adapter was utilized during the
execution of the workload (612a), and so on. Monitoring (704) the
consumption of various resources by the workload (612a) in one or
more runs of the workload (612a) may therefore provide information
regarding the actual usage of system resources when executing the
workload (612a), thereby enabling more informed decisions to be
made regarding the deployment of the workload (612a).
[0068] In the example of FIG. 7, selecting (618) a set of nodes
(602a-602e) having attributes that meet the specific resource
requirements and arranged to meet the required geometry includes
selecting (706) a plurality of candidate sets of nodes. In the
example of FIG. 7, selecting (706) a plurality of sets of nodes
(602a-602e) provides a plurality of candidate sets of nodes
(602a-602e) upon which a workload (612a-612c) may ultimately be
deployed. In the example of FIG. 7, selecting (706) a plurality of
sets of nodes may be carried out, for example, by including every
possible permutation of node sets in the candidate sets of nodes,
by including only sets of nodes with particular attributes in the
candidate sets of nodes, and so on.
[0069] In the example of FIG. 7, selecting (618) a set of nodes
(602a-602e) having attributes that meet the specific resource
requirements and arranged to meet the required geometry also
includes assigning (708) to each candidate set of nodes a score,
the score being a representation of the degree to which the
attributes of the nodes of the candidate set meet the resource
requirements of the workload (612a) and the geometry requirements
of the workload (612a). In the example of FIG. 7, each score may be
calculated in a variety of ways. For example, each score may be
calculated as a percentage of the specific resource requirements
that are satisfied by a particular candidate set of nodes, as a
weighted score in which particular resource requirements of high
importance are given a higher value than particular resource
requirements of high importance, and so on. In the example of FIG.
7, selecting (618) a set of nodes (602a-602e) having attributes
that meet the specific resource requirements and arranged to meet
the required geometry may therefore include selecting (710) the set
of nodes having the best score.
[0070] For further explanation, FIG. 8 sets forth a flow chart
illustrating a further exemplary method for optimizing the
deployment of a workload on a distributed processing system (601)
according to embodiments of the present invention. The example of
FIG. 8 is similar to the example of FIG. 6 as it also includes:
[0071] profiling (604) during operations on the distributed
processing system (601) attributes (606) of the nodes (602a-602e)
of the distributed processing system (601), [0072] selecting (608)
a workload (612a) for deployment on a subset (622) of the nodes
(602a-602e) of the distributed processing system (601), [0073]
determining (614) specific resource requirements for the workload
(612a) to be deployed, [0074] determining (616) a required geometry
of the nodes (602a-602e) to run the workload (612a), [0075]
selecting (618) a set of nodes (602a-602e) having attributes that
meet the specific resource requirements and arranged to meet the
required geometry, and [0076] deploying (620) the workload (612a)
on the selected nodes (602a-602e)
[0077] In the example of FIG. 8, profiling (604) during operations
on the distributed processing system (601) attributes (606) of the
nodes (602a-602e) of the distributed processing system (601)
includes profiling (802) the attributes of a set of nodes during a
previous run of the workload (612a). As described above with
reference to FIG. 6, profiling (604) attributes (606) of the nodes
(602a-602e) may be carried out, for example, by recording
performance metrics of the nodes (602a-602e). For example,
performance metrics may be recorded and compiled such that a
profile regarding the performance attributes of the nodes
(602a-602e) can be created that details the amount of time a node
(602a-602e) took to carry out a floating point operation, the
amount of time a node (602a-602e) took to carry out an I/O
operation, the amount of time a node (602a-602e) took to carry out
a data communications operation, and so on. In the example of FIG.
8, however, profiling (604) during operations on the distributed
processing system (601) attributes (606) of the nodes (602a-602e)
of the distributed processing system (601) includes profiling (802)
the attributes of a set of nodes during a previous run of the
workload (612a). That is, the workload (612a) may be executed and
performance metrics may be recorded while the workload (612a) is
being executed. These recorded performance metrics may be used at a
later time such that profiling (604) attributes (606) of the nodes
(602a-602e) takes into account the recorded performance metrics,
even if the workload (612a), or other workloads (612b, 612c) have
been executed in the interim.
[0078] In the example of FIG. 8, selecting (608) a workload (612a)
for deployment on a subset (622) of the nodes (602a-602e) of the
distributed processing system (601) includes selecting (804) a set
of nodes that are different than those used in the previous run of
the workload (612a). In the example of FIG. 8, profiling (802) the
attributes of a set of nodes during a previous run of the workload
(612a) can include recording information identifying the particular
set of nodes that executed the workload (612a) during the previous
run. In such an example, selecting (608) a workload (612a) for
deployment on a subset (622) of the nodes (602a-602e) of the
distributed processing system (601) can include selecting some
combination of nodes other than the combination of nodes that
executed the workload (612a) during the previous run.
[0079] In the example of FIG. 8, deploying (620) the workload
(612a) on the selected nodes (602a-602e) includes suggesting (806)
the set of nodes that are different than those used in the previous
run of the workload (612a) for the next run of the workload (612a).
In the example of FIG. 8, suggesting (806) the set of nodes that
are different than those used in the previous run of the workload
(612a) for the next run of the workload (612a) may be carried out,
for example, by delivering a prompt to the user suggesting the set
of nodes that are different than those used in the previous run of
the workload (612a). In such an example, the user may choose, via
the prompt, to use the set of nodes that are different than those
used in the previous run of the workload (612a) for the next run of
the workload (612a).
[0080] The example of FIG. 8 also includes identifying (808) in
dependence upon the attributes of the nodes (602a-602e) of the
distributed processing system (601), components to be replaced and
suggesting the replacement of the components. In the example of
FIG. 8, identifying (808) components to be replaced may be carried
out, for example, by identifying nodes (602a-602e) that are
malfunctioning, by identifying nodes (602a-602e) with attributes
that are low priority attributes for executing a particular
workload (612a-612c), and so on. In the example of FIG. 8,
suggesting the replacement of the components may be carried out,
for example, by delivering a prompt to a user such as a system
administrator that identifies the component to be replaced and
suggests a replacement component that is more fit for executing a
particular workload (612a-612c).
[0081] For further explanation, FIG. 9 sets forth a flow chart
illustrating a further exemplary method for optimizing the
deployment of a workload on a distributed processing system (601)
according to embodiments of the present invention. The example of
FIG. 9 is similar to the example of FIG. 6 as it also includes:
[0082] profiling (604) during operations on the distributed
processing system (601) attributes (606) of the nodes (602a-602e)
of the distributed processing system (601), [0083] selecting (608)
a workload (612a) for deployment on a subset (622) of the nodes
(602a-602e) of the distributed processing system (601), [0084]
determining (614) specific resource requirements for the workload
(612a) to be deployed, [0085] determining (616) a required geometry
of the nodes (602a-602e) to run the workload (612a), [0086]
selecting (618) a set of nodes (602a-602e) having attributes that
meet the specific resource requirements and arranged to meet the
required geometry, and [0087] deploying (620) the workload (612a)
on the selected nodes (602a-602e)
[0088] In the example of FIG. 9, profiling (604) during operations
on the distributed processing system (601) attributes (606) of the
nodes (602a-602e) of the distributed processing system (601) may
include running (902) a system exerciser (910) on the distributed
processing system (601). In the example of FIG. 9, the system
exerciser (910) includes operations to test the attributes of the
nodes (602a-602e). In the example of FIG. 9, the system exerciser
(910) may be embodied as automated computing machinery such as a
module of computer program instructions executing on computer
hardware. The system exerciser (910) may include various
operations, embodied as computer program instructions, which test
the attributes of the nodes (602a-602e). The system exerciser (910)
may test the attributes of the nodes (602a-602e), for example, by
executing floating point operations, data communications
operations, memory access operations, and the like to determine how
quickly each of the nodes (602a-602e) may carry out the operations
such that a processing profile may be compiled for each of the
nodes (602a-602e).
[0089] In the example of FIG. 9, profiling (604) during operations
on the distributed processing system (601) attributes (606) of the
nodes (602a-602e) of the distributed processing system (601) may
also include recording (904) the resultant performance of the
attributes of the nodes (602a-602e) in response to the system
exerciser (910). In the example of FIG. 9, recording (904) the
resultant performance of the attributes of the nodes (602a-602e) in
response to the system exerciser (910) may be carried out, for
example, by recording (904) the resultant performance of the
attributes of the nodes (602a-602e) in a special purpose node
attribute table that includes measured performance metrics for each
of the nodes (602a-602e).
[0090] In the example of FIG. 9, selecting (618) a set of nodes
(602a-602e) having attributes that meet the specific resource
requirements and arranged to meet the required geometry includes
suggesting (908) an initial set of nodes for deploying the workload
(612a). In the example of FIG. 9, suggesting (908) an initial set
of nodes for deploying the workload (612a) may be carried out, for
example, by sending a prompt to a user such as a system
administrator, by simply deploying the selected (618) set of nodes
(602a-602e), and so on.
[0091] In the example of FIG. 9, profiling (604) during operations
on the distributed processing system (601) attributes (606) of the
nodes (602a-602e) of the distributed processing system (601) may
include storing (906) in a database an identification of the nodes
and the specific attributes of the nodes. Storing (906) an
identification of the nodes and the specific attributes of the
nodes in a database may be carried out, for example, by writing the
attributes of the nodes to the database upon completion of a
workload (612a-612c), by writing the attributes of the nodes to the
database at predetermined intervals such as a predetermined
interval of time or a predetermined number of workload executions,
by writing the attributes of the nodes to the database when total
system utilization drops below a predetermined threshold, and so
on.
[0092] As will be appreciated by one skilled in the art, aspects of
the present invention may be embodied as a system, method or
computer program product. Accordingly, aspects of the present
invention may take the form of an entirely hardware embodiment, an
entirely software embodiment (including firmware, resident
software, micro-code, etc.) or an embodiment combining software and
hardware aspects that may all generally be referred to herein as a
"circuit," "module" or "system." Furthermore, aspects of the
present invention may take the form of a computer program product
embodied in one or more computer readable medium(s) having computer
readable program code embodied thereon.
[0093] Any combination of one or more computer readable medium(s)
may be utilized. The computer readable medium may be a computer
readable signal medium or a computer readable storage medium. A
computer readable storage medium may be, for example, but not
limited to, an electronic, magnetic, optical, electromagnetic,
infrared, or semiconductor system, apparatus, or device, or any
suitable combination of the foregoing. More specific examples (a
non-exhaustive list) of the computer readable storage medium would
include the following: an electrical connection having one or more
wires, a portable computer diskette, a hard disk, a random access
memory (RAM), a read-only memory (ROM), an erasable programmable
read-only memory (EPROM or Flash memory), an optical fiber, a
portable compact disc read-only memory (CD-ROM), an optical storage
device, a magnetic storage device, or any suitable combination of
the foregoing. In the context of this document, a computer readable
storage medium may be any tangible medium that can contain, or
store a program for use by or in connection with an instruction
execution system, apparatus, or device.
[0094] A computer readable signal medium may include a propagated
data signal with computer readable program code embodied therein,
for example, in baseband or as part of a carrier wave. Such a
propagated signal may take any of a variety of forms, including,
but not limited to, electro-magnetic, optical, or any suitable
combination thereof. A computer readable signal medium may be any
computer readable medium that is not a computer readable storage
medium and that can communicate, propagate, or transport a program
for use by or in connection with an instruction execution system,
apparatus, or device.
[0095] Program code embodied on a computer readable medium may be
transmitted using any appropriate medium, including but not limited
to wireless, wireline, optical fiber cable, RF, etc., or any
suitable combination of the foregoing.
[0096] Computer program code for carrying out operations for
aspects of the present invention may be written in any combination
of one or more programming languages, including an object oriented
programming language such as Java, Smalltalk, C++ or the like and
conventional procedural programming languages, such as the "C"
programming language or similar programming languages. The program
code may execute entirely on the user's computer, partly on the
user's computer, as a stand-alone software package, partly on the
user's computer and partly on a remote computer or entirely on the
remote computer or server. In the latter scenario, the remote
computer may be connected to the user's computer through any type
of network, including a local area network (LAN) or a wide area
network (WAN), or the connection may be made to an external
computer (for example, through the Internet using an Internet
Service Provider).
[0097] Aspects of the present invention are described above with
reference to flowchart illustrations and/or block diagrams of
methods, apparatus (systems) and computer program products
according to embodiments of the invention. It will be understood
that each block of the flowchart illustrations and/or block
diagrams, and combinations of blocks in the flowchart illustrations
and/or block diagrams, can be implemented by computer program
instructions. These computer program instructions may be provided
to a processor of a general purpose computer, special purpose
computer, or other programmable data processing apparatus to
produce a machine, such that the instructions, which execute via
the processor of the computer or other programmable data processing
apparatus, create means for implementing the functions/acts
specified in the flowchart and/or block diagram block or
blocks.
[0098] These computer program instructions may also be stored in a
computer readable medium that can direct a computer, other
programmable data processing apparatus, or other devices to
function in a particular manner, such that the instructions stored
in the computer readable medium produce an article of manufacture
including instructions which implement the function/act specified
in the flowchart and/or block diagram block or blocks.
[0099] The computer program instructions may also be loaded onto a
computer, other programmable data processing apparatus, or other
devices to cause a series of operational steps to be performed on
the computer, other programmable apparatus or other devices to
produce a computer implemented process such that the instructions
which execute on the computer or other programmable apparatus
provide processes for implementing the functions/acts specified in
the flowchart and/or block diagram block or blocks.
[0100] The flowchart and block diagrams in the Figures illustrate
the architecture, functionality, and operation of possible
implementations of systems, methods and computer program products
according to various embodiments of the present invention. In this
regard, each block in the flowchart or block diagrams may represent
a module, segment, or portion of code, which comprises one or more
executable instructions for implementing the specified logical
function(s). It should also be noted that, in some alternative
implementations, the functions noted in the block may occur out of
the order noted in the figures. For example, two blocks shown in
succession may, in fact, be executed substantially concurrently, or
the blocks may sometimes be executed in the reverse order,
depending upon the functionality involved. It will also be noted
that each block of the block diagrams and/or flowchart
illustration, and combinations of blocks in the block diagrams
and/or flowchart illustration, can be implemented by special
purpose hardware-based systems that perform the specified functions
or acts, or combinations of special purpose hardware and computer
instructions.
[0101] It will be understood from the foregoing description that
modifications and changes may be made in various embodiments of the
present invention without departing from its true spirit. The
descriptions in this specification are for purposes of illustration
only and are not to be construed in a limiting sense. The scope of
the present invention is limited only by the language of the
following claims.
* * * * *